Skip to content
ChatbotIQ

What Are Knowledge Gaps?

A knowledge gap is a topic where users are asking questions but your knowledge base doesn’t have good answers. ChatbotIQ detects these automatically and tells you exactly what content to add.

How gap detection works

Section titled “How gap detection works”

Every time your bot answers a question, ChatbotIQ tracks how well it was able to answer — from strong matches to no relevant content found at all.

Periodically, ChatbotIQ analyzes all the poorly-answered queries and groups related questions into topic clusters. Each cluster represents a knowledge gap — an area where your content is missing or insufficient.

Types of knowledge gaps

Section titled “Types of knowledge gaps”

ChatbotIQ classifies each gap into one of three types:

Content gap

Section titled “Content gap”

Your product or service covers this topic, but your knowledge base is missing documentation about it. Action: add the missing content.

For example, if users keep asking “How do I export data?” but you haven’t documented the export feature, that’s a content gap.

Product gap

Section titled “Product gap”

Users are asking about a feature or capability that your product doesn’t have. Action: note as product feedback.

For example, if users ask “Can I integrate with Slack?” and you don’t offer Slack integration, that’s a product gap. You might add a Q&A entry explaining this, or share the feedback with your product team.

Out of scope

Section titled “Out of scope”

The question isn’t related to your product or bot’s purpose. Action: consider refining your bot’s system prompt to explain its scope.

For example, if users ask “What’s the weather today?” to your documentation bot, that’s out of scope.

What makes gap detection valuable

Section titled “What makes gap detection valuable”

No guesswork

Section titled “No guesswork”

Instead of wondering what content to add next, you have a data-driven list of topics sorted by how many users asked about them.

Real user language

Section titled “Real user language”

The sample queries show you the exact words users typed. This helps you write content that matches how people actually search — not just how your team would phrase things.

Prioritization

Section titled “Prioritization”

Gaps are ranked by query count and decline rate. A gap that 50 users hit is more urgent than one that 3 users hit.

Progress tracking

Section titled “Progress tracking”

You can mark gaps as resolved to track your improvement over time.

The improvement cycle

Section titled “The improvement cycle”
  1. Launch your bot and let it handle real conversations.
  2. Check Analytics after a week or two. Look at the knowledge gaps section.
  3. Address content gaps by adding or updating your knowledge sources.
  4. Refresh your sources so the bot picks up the new content.
  5. Check again in another week. The gaps should shrink.

Over time, your bot gets smarter — not because the AI model improves, but because your knowledge base covers more of what users actually ask about.

Section titled “Related”